The attraction of using artificial intelligence (AI) to support judicial decision-making in the administrative context is obvious, but the considerations used for judicial review of algorithmic decision-making (ADM) ought also to be applied to judicial use of ADM in order to ensure its fair and optimal use. This chapter focuses on the UK experience to argue that five factors should be considered: (1) the potential for a particular area to be technologically justiciable; (2) the definition of a fair procedure and the need to choose the model or form of automation to fit the particular purpose and context of the system; (3) the need for gisting, wherever it is important to provide contestability of the system; (4) the iterative adoption of technology at a macro level, with a proportionate right to individual-level accuracy at the micro level; and (5) the need for safeguards on the use of data from one area in another. The availability of these insights in public law has the capacity to inform our choices, not only in the administrative justice context, but also across the board.

Judicial independence is an essential part of democracies, based on the division of powers, rule of law, and respect for fundamental rights. In its most simplified version, judicial independence relies on the freedom from (and resilience to) the external and internal influences and pressures that the courts as institutions and judges as justice professionals are constantly subject to. Introducing AI into the judicial system could impact the judicial independence from a wide spectrum of angles: judicial independence can be compromised and shaped by the AI systems, in particular if these systems have been developed by private sector and/or designed by legislative or executive powers. Furthermore, AI systems can do this in much less perceptible ways that are difficult to detect and complicated to prove, for instance through the experts that courts rely on when the case requires specific knowledge or expertise. This chapter focuses on identifying these threats and addressing them in a constructive and solution-oriented manner, without compromising the potential of AI for the justice system.

Integrating algorithmic tools into judicial systems prompts critical questions on public trust, due process, and fairness, alongside inherent risks of the pursuit of ‘technical fix’. In response to growing demands for transparency and consistency, Taiwan has introduced algorithmic and AI-powered sentencing tools, representing significant steps toward reforming sentencing practices and improving judicial accountability. However, their implementation has encountered formidable challenges, including low adoption rates, judicial misunderstandings, algorithmic biases, and insufficient regulatory frameworks. This chapter explores these issues within Taiwan’s historical and legal context, providing an in-depth analysis of empirical data and judicial practices. By situating Taiwan’s experience within the global discourse on AI in judicial systems, the chapter illuminates the complexities of integrating AI into a civil law tradition while striving to maintain judicial independence. Taiwan’s approach offers insights for jurisdictions worldwide, contributing to broader discussions on leveraging AI to enhance justice without compromising foundational legal principles and values.

The increasing reliance on algorithmic and AI systems by judges is reshaping the judiciary and its way of working in numerous ways. One aspect that has remain under- examined is how the judicial duty to state reasons may be affected. This refers to the obligation of judges to provide reasons whenever they rule in a case. In fact, the duty constitutes an essential component of the rule of law and the right to a fair trial, and fulfils important normative goals, such as legitimacy, transparency, and accountability of the judicial decision-making process. The chapter therefore first provides a concise conceptualisation of the duty, including its underlying normative goals. It then examines how and to what extent the judicial duty to state reasons can be impacted whenever judges rely on AI systems, focusing on the impact of such systems on the underlying normative goals of the duty. The chapter concludes with some reflections on how the duty can be safeguarded in the age of automation.

This chapter focuses on the legal framework for the use of AI in courts in Croatia and Slovenia, which results from their legal traditions as well as their membership in the Council of Europe and the EU. It also aims to discuss AI systems, either operational or in development, in both countries, and to evaluate their impact on fundamental rights and ethics. The findings demonstrate that while both countries experience a slow but gradual introduction of AI initiatives, in Slovenia this is happening without pre-existing or rigorous regulatory oversight.

This chapter examines how artificial intelligence (AI) can address inefficiencies in India’s judicial system, focusing on Protection of Children from Sexual Offences (POCSO) cases. Analysis of 220,000 cases reveals significant regional disparities in processing and outcomes, reflecting broader systemic challenges. Despite digital infrastructure investments, we identify a disconnect between data collection and data- driven decision-making. We propose an AI-powered dashboard to provide real-time case tracking, identify bottlenecks, and improve resource allocation. While implementation faces challenges related to data quality and privacy, successful deployment could serve as a model for judicial reform in India and globally.

Legal futurists have urged judiciaries to experiment with automated legal systems. However, for many aspects of legal systems, there is a common sense that their translation into computation would be inappropriate. The simultaneous malleability of legal systems and prevalence of constitutive practices within them, should lead to a two-level consideration of (1) what aspects of a liberal legal order are crucial, and (2) for those that are crucial, what is lost when that aspect is either partially or fully automated. In legal decision-making, some patterns of action are merely instrumental to achieving ends, while others are essential, or constitutive: the activity should no longer even be considered part of a liberal legal order when the practice ceases. Administrative processes that are simply incidental and instrumental to the legitimate resolution of a case are well primed for automation. Other practices are essential and intrinsically important, and properly resist being converted into machine-readable code. Distinctions between incidental and constitutive, or instrumentally and intrinsically important, aspects of law, should both bound and guide legal automation.

Can AI adjudicative tools in principle better enable us to achieve the rule of law by replacing judges? This chapter argues that answers to this question have been excessively focused on ‘output’ dimensions of the rule of law – such as conformity of decisions with the applicable law – at the expense of vital ‘process’ considerations such as explainability, answerability, and reciprocity. These process considerations do not by themselves warrant the conclusion that AI adjudicative tools can never, in any context, properly replace human judges. But they help bring out the complexity of the issues – and the potential costs – that are involved in this domain.

Artificial intelligence (AI) has started to make its way into Spanish court practice, especially in criminal justice. Furthermore, this trend has been accompanied by two new regulations: the EU AI Act, the world’s first comprehensive law on the topic, and the Spanish Royal Decree-Law No. 6 of 2023. At present, there are already several AI-based tools used by Spanish courts and the application of them proves highly beneficial, in particular in certain areas of criminal justice. Nevertheless, AI use can pose serious problems related to conflict with different fundamental rights of the accused. Therefore, its use should be considered with great caution.

Artificial intelligence (AI) increasingly intersects with judicial processes, raising new challenges for courts and judges. One significant concern linked to this development is the ability of judges and court personnel to understand, evaluate, and critically engage with AI systems. The EU Artificial Intelligence Act adopted in 2024 addresses this directly, requiring public bodies using AI to ensure their staff possess a ‘sufficient level of AI literacy’. This chapter argues that enhancing AI competence among judges and court personnel is essential to safeguarding the right to a fair trial, legal certainty, and the rule of law in an increasingly digitalised legal environment. After providing a brief overview of AI literacy obligations in the EU AI Act, the chapter offers insights into how national judicial training institutions could integrate AI literacy into their curricula.

The integration of artificial intelligence (AI) into judicial decision-making presents both opportunities and challenges, particularly in balancing legal certainty and judicial discretion. While AI-driven tools are designed to enhance consistency and efficiency, their growing influence may subtly reshape judicial reasoning, potentially narrowing judicial discretion. This chapter examines how algorithmic recommendations – rather than merely assisting adjudication – can become dominant reference points, steering judicial outcomes toward standardisation over case-specific interpretation. Drawing on empirical psychological research, behavioural law, and economics, and the works of Richard Posner, Aharon Barak, and other legal theorists, the chapter explores the psychological mechanisms underlying this shift, particularly phenomena known as ‘automation bias’ and the ‘anchoring effect’, which may unconsciously influence judicial decision-making. The analysis highlights these psychological and structural challenges, inviting reflection on how AI-driven legal certainty impacts judicial discretion and the space for individualised legal reasoning in modern adjudication.

This chapter focuses on AI and its impact on transparency in judicial decision-making. Transparency is one of the core values of the rule of law, and essential for maintaining the trust and accountability of the judiciary and justice system as a whole. Drawing upon semi-structured expert interviews with members of judiciary and legal profession, case law and real-life examples of AI tools, the chapter considers four questions: why transparency matters in the context of judicial decision-making; the information that judges must have and communicate to satisfy the demands of transparency; whether they have access to this information; and, if not, what we might do about this deficit. We argue that two complementary solutions can strengthen judicial transparency in the age of AI: (1) a regulatory framework that mandates disclosure of specific information pertaining to the code and variables used in AI tools; and (2) robust use of the due process duty to provide adequate reasons for a judicial decision that depends upon the output of a predictive tool. These steps are essential to reconcile judicial use of AI with the need for transparency, as a foundational aspect of justice and rule of law.

International human rights courts and treaty bodies are increasingly turning to automated decision-making (ADM) technologies to expedite and improve their review of individual complaints. These tribunals have yet to consider many of the legal, normative, and practical issues raised by the use of different types of automation technologies for these purposes. This chapter offers an initial assessment of the benefits and challenges of introducing ADM into international human rights adjudication. We weigh up the benefits of introducing these tools to improve international human rights adjudication – which include greater speed and efficiency in processing and sorting cases, identifying patterns in jurisprudence, and enabling judges and staff to focus on more complex responsibilities – against two types of cognitive biases – biases inherent in the datasets on which ADM is trained, and biases arising from interactions between humans and machines. We also introduce a framework for enhancing the accountability of ADM tools that mitigates the potential harms caused by automation technologies in this context.

Say an AI program passes a Turing test because it can converse in a way indistinguishable from a human. And say that its developers can then teach it to converse – and even present an extended persuasive argument – in a way indistinguishable from the sort of human we call a ‘lawyer’. The program could thus become an AI brief-writer, capable of regularly winning brief-writing competitions against human lawyers. If and when that happens, this chapter argues, the same technology can be used to create AI judges, judges that we should accept as no less reliable than human judges, and more cost-effective. If the software can create persuasive opinions, capable of regularly winning opinion-writing competitions against human judges, we should accept it as a judge, even if the opinions do not stem from human judgment.

This chapter provides a comprehensive overview of the history and developments of AI in courts. In particular, through the lens of legal informatics, we explore four phases in the development and evolution of AI in courts: judicial information retrieval, human-made models of judicial reasoning, machine learning for judicial prediction, and large language models for courts. For each of these, we explore the opportunities and challenges in their implementation and adoption within the judicial system.